The impact of integrated nutrition-sensitive interventions on nutrition and health of children and women in rural Tanzania: Study protocol for a cluster-randomized controlled trial

Background: Nutrition-sensitive interventions such as homestead production of diverse, nutrient-rich foods, coupled with behavior change communication, may have positive effects on the nutritional status and health of rural households engaged in agriculture, particularly among women and young children. Engagement of agriculture and health extension workers in these communities may be an effective way of delivering nutrition-sensitive interventions given the dearth of trained health care providers in many developing countries. This study aims to assess the effects of integrated homestead food production, food consumption and women’s empowerment interventions using a multi-sectoral approach on women’s and child’s health and nutrition. Methods: This is a cluster-randomized community-based prospective study set in Rufiji district, a rural area in Eastern Tanzania. Ten randomly selected villages within the Rufiji Health and Demographic Surveillance Site (HDSS) in Eastern-Tanzania were paired and randomly assigned to the intervention or control arm. The Rufiji HDSS dataset was used to randomly sample households with women of reproductive age and children 6-36 months. The intervention includes provision of small agricultural inputs, garden training support, and nutrition and health counseling. This is delivered by community health workers and agriculture extension workers through home visits and farmer field schools. There are three time points for data collection: baseline, midline, and endline. Primary outcomes are women’s and children’s dietary diversity, maternal and child anemia status and growth (child stunting, child wasting, women’s BMI, and women and child hemoglobin). Discussion: This integrated agriculture and nutrition intervention among rural farming households is a simple and innovative solution that may contribute to the reduction of undernutrition and disease burden among women and children in developing countries. Engaging already existing workforce in the community may facilitate sustainability of the intervention package. Trial registration: ClinicalTrials.gov NCT03311698, Retrospectively registered on October 17, 2017.

The study is implemented in Rufiji district, Tanzania. Rufiji is one of six districts in the Coast Region of Tanzania, and has a population size of 217,274. It has 94 registered villages and 385 hamlets. The study is operated within the Rufiji Health and Demographic Surveillance Site (RHDSS) where about 20% of households have at least one child under the age of 3 years [18]. The prevalence of stunting and anemia among children under 5 years in Rufiji is 35 and 61% respectively [19]. Its demographic characteristics are representative of rural Tanzania, most adults are subsistence farmers who cultivate staples like maize, rice, cassava, and millet Fig. 1. Map of the study area in Rufiji, Tanzania This study engages existing Agriculture Extension Workers (AEWs), Livestock Extension Workers (LEWs) and Community Health Workers (CHWs) to deliver the interventions. AEWs and LEWs are employed by the government to train and support farmers in agriculture and animal husbandry and best practices for increased productivity. CHWs are an existing cadre in many rural settings of Tanzania, but have no formal wages. The project therefore takes advantage of existing cadres of workers who are familiar with their communities. Typically, each AEW, LEW and CHW works in one village with approximately 400 households. They conduct regular home visits among all households to which they have been assigned providing basic health education, promoting preventive health care services, and supporting best practices in agriculture and livestock production. All three forms of home-based services; health, agriculture and livestock exist in all Rufiji villages. This is a 36-month long community-based prospective study carried out in 10 villages. A cluster-randomized design is used to evaluate the effects of the intervention. Data collection take place at three time points, at baseline, midline, and endline. Study activities are summarized in Fig. 2. Study activities over study period Eligibility criteria for households participating in the study are: (i) having a woman of reproductive age (18–49 years) and at least one child 6–36 months of age, (ii) having access to a plot of land or containers where vegetables could be grown, and (iii) provided informed consent. The unit of randomization is a village. Ten villages within the Rufiji HDSS were randomly selected and matched based on location, proximity to water, and population size. One village in each matched-pair was randomly assigned to receive the intervention (five villages) and the other to the control group (five villages). The Rufiji HDSS dataset was used in random sampling of households with at least one woman of reproductive age and at least one child 6–36 months of age in all study villages. We considered a proportional-based representation of households based village size in the random selection of households. Thus, a total of 1006 households were enrolled with a ratio of 1:1 in intervention and control villages (Table 1). Enrolment was initiated on July 2016 and the endline survey is planned for July 2019 Fig. 2. Sample frame of randomized study villages and households The sample size was pre-determined by the size of HDSS and the logically feasible time frame of 2 years. The number of eligible households which could be enrolled was estimated before (as 1000) to be sufficient for pilot implementation of integrated nutrition-sensitive interventions but no formal sample size calculation was performed. Participating households in each study arm were screened, consented, and enrolled. Research assistants (RA) visited and screened sampled households for each study village. RAs introduced the study to the household member and verified participant’s inclusion criteria. For each eligible household, one woman and one child were enrolled in the study after providing informed consent. In case the household has more than one eligible member, simple random sampling technique (lottery method) within each household was used to select one woman. The same technique was used to select one child 6–36 months of age. RAs then proceeded with administration of electronic baseline data collection. Households in the intervention arm receive: (1) interventions to promote homestead food production and increase agricultural production and food diversity, (2) nutritional counselling, including prevention and management of child malnutrition with a focus on the first 1000 days, locally adapted instructions on the mix and quantity of food suitable for children ages 6–36 months, (3) a health-focused intervention, including information on micronutrient supplementation, integrated management of child illnesses, and safe water, sanitation and hygiene practices (WASH). Households enrolled in the intervention also receive agricultural inputs such as seeds, fertilizer, and watering cans. Local Agricultural Officers identified six varieties of crops (Africa eggplant, amaranthus, spinach, tomato, okra, and Chinese cabbage) for the study and recommended specific varieties for each intervention village based on growing conditions and village preference. Each participating household in intervention villages receive three varieties of seeds at least three times during the study period. The behavior change component of the intervention includes information on cultivation of nutrient-rich crop varieties, best practices for home gardening, safe water, sanitation, and hygiene practices (WASH), breastfeeding and complementary feeding, dietary intake and nutrition, and other basic public health messages. AEWs, LEWs and CHWs are trained on topics related to cultivation of nutrient-rich crop varieties and provision of basic public health services to households. This includes encouraging and monitoring the consumption of iron and folate supplements during pregnancy; use of zinc oral rehydration solution for management of diarrhea; promotion of hygienic and sanitary practices at the household level, promotion of preventive health behaviors such as immunization for children; and referral of sick and malnourished children to local health centers for appropriate care. There are two delivery mechanisms for the intervention: farmer field schools and households visits. Household visits are conducted approximately every 2 weeks. Households receive separate visits from an AEW/LEW or CHW on a rotating schedule. Field schools are also held approximately every 2 weeks and attended by all three types of extension workers. In collaboration with village and hamlet leaders, AEWs identified one participating household farm in each of four to five hamlets per village to be used as Farmer Field Schools (FFS). Participants within a given hamlet were invited to partake in FFSs. Farmer field schools are open to the public as they also existed in these communities before the intervention took place. Households participating in the intervention are highly encouraged to attend by extension workers. The number of participants attending a FFS typically ranges from 10 to 15, determined by the number of enrolled households per hamlet. Collaboration with village and hamlet leaders ensure that the selected individuals are valued and trusted community members interested in leading fellow farmers. The FFS is designed to (i) illustrate the benefits of innovative agricultural practices to the community, (ii) improve community knowledge about nutritious crops available in the district, (iii) reinforce messages provided by during household visits, (iv) offer a forum for discussion and collaboration between farmers, and (v) empower women with knowledge and skills for increased decision-making at the household level. Model farmers use the demonstration farm as a classroom and hold organized “classes” with other women farmers in their hamlet to illustrate how to become a more productive farmer. The design of FFS enables AEWs to maintain close connections with model farmers, providing them with information about new agricultural technologies and practices as they become available in the district. AEWs, LEWs and CHWs implement a system of household visits with all enrolled smallholder farming households in the community. All cadres receive basic nutritional, agricultural and maternal and child health training to ensure messages act to reinforce and complement each other. The transmission of the same messages through multiple channels aims to increase their credibility and the probability of longer term adoption of recommended behavioral change. One-on-one meetings with AEWs, LEWs and CHWs through home visit are held biweekly for the purposes of providing maximum amount of information in a feasible manner. Household visits and FFSs are staggered so that farmers receiving the intervention are continually exposed to information through one of these channels. Extension workers are continuously monitored by the study field manager to sure that the above two main approaches are delivered routinely to participants as outlined in an intervention manual developed by the research team in collaboration with local officials. A list of randomly selected participating households in each intervention village is generated every week and visited by the field manager to conduct monitoring and evaluation of the intervention. Data are collected by trained interviewers using surveys administered with electronic tablets at three time points, once at baseline, once 12 months after intervention implementation, and once approximately 36 months after intervention implementation. Agricultural practices and primary outcomes are also measured in a random sample of non-participant households in the intervention villages which may assist to inform spillover effect or spread of the intervention. Anthropometric measurements of women and young children take place at three time points. Hemoglobin for women and young children is measured among a subset of the study population at these time points. Questionnaires developed by the research team include socio-demographic and economic data, self-reported recent history of illness, infant and young child feeding, dietary intake, physical activity, agriculture production, food security, and women’s empowerment. Agriculture production data include the area of specified crop production, production quantities, sales of produce and income realized, purchase and use of agriculture inputs, and other operational expenses. Dietary intake is assessed using a food frequency questionnaire. Qualitative methods are used to assess the reach of the intervention, opportunities and challenges. Focus group discussions will be conducted with study participants and non-study participants in intervention villages. Key informant interviews will also be conducted with health care providers, extension workers, and village and district officials. Interviews and focus groups aim to ascertain participants’ experiences with the intervention and suggestions for improvement in the delivery and sustainability of the intervention. In addition, the interviews and discussion will also identify bottlenecks and sources of improvement or adaption of homestead agriculture and nutrition interventions in the community. The primary outcome for the study is women’s and children’s dietary diversity (number of food groups consumed out of 10) [20, 21]. Secondary outcomes include women’s and children’s anemia status (hemoglobin [Hb] levels < 12 g/dl for non-pregnant women and Hb < 11 g/dl for pregnant women, and Hb < 11 g/dl for children); child growth (weight for age z-score [WAZ], weight for height z-score [WHZ], height for age z-score [HAZ], and mid upper arm circumference [MUAC]); women’s growth (body mass index, and MUAC); early childhood development (Caregiver Reported Early Development Index [CREDI]) [22]; and reach and extent of intervention-promoted practices and behaviors (home gardening and food consumption). Data will be analyzed based on an intent-to-treat strategy. Data will be classified into three main categories; (i) the primary outcome variables including women and children dietary diversity, (ii) secondary outcomes including women’s and children’s growth and anemia status; and (iii) other covariates including socio-demographic characteristics, village, age, gender, health status, WASH practices, and crop production and practices, and livestock ownership. In addition, standard tools to assess food security (food reserve availability index), dietary diversity in women of reproductive age and children, women empowerment in agriculture index, water and sanitation (WASH) practices as well as household asset index will be incorporated, and aggregate data used to predict vulnerable households. Height (or length) and weight will be standardized based on 2006 WHO child growth standards to create HAZ, WAZ and WHZ scores. Biological implausible values based on WHO recommended cutoffs at 6 SD will be eliminated. Generalized linear mixed models with random intercepts and robust standard errors will be used to assess for continuous repeated primary outcomes. Generalized linear mixed models with random intercepts and robust standard errors will be used to assess repeated primary outcomes. Enrolment into the study began in July 2016 and completed in November 2017. The intervention period is August 2016 through July 2019. The study is ongoing.

N/A

• Dima Health